What's new
What's new

CNC Lathe cutting rough arcs, need help troubleshooting.

Sinij kot

Plastic
Joined
Jul 30, 2009
Location
Albany NY
Hello, I'm working on 1995 Supermax TC-2 cnc lathe. I recently purchased this machine after it has been sitting for around 5 years and the previous owner working on it on and off trying to repair it. It has Mitsubishi Meldas L3 control which I've been getting used to. After repairing few things and getting parameters set correctly I'm doing test parts to see how the lathe performs and ironing out post processor.
The issue I ran into has to do with surface finish on the arc. I'm cutting a part out of 6061 which is short piece of tube ~0.75" long 2.5" o.d. 2"i.d. and has a ball nose end. I'm roughing with CNMG432 for AL and finishing with [FONT=arial, sans-serif]CCGT 32.51 for AL. The finish pass is 0.01" DOC and 0.0015" IPR. The finish looks good on the straight portion of the tool-path but when the radius is cut the it is not smooth and has steps.[/FONT]
[FONT=arial, sans-serif]I checked the G-code and it is commanding G03 X.. Y.. R.. so I know it is not cutting in segments. When I first noticed this problem I thought that the CNMG432 was part of the cause but going to CCGT 32.51 still produces similar profile on the radius and improved finish on the straight section.[/FONT]
[FONT=arial, sans-serif]I think this might be a problem with the machine it self. I decided to take a closer look at the part under the microscope and see if I can learn anything. Below are few pictures from the microscope:
[/FONT]finish issues6.jpgfinish issues5.jpgfinish issues4.jpgfinish issues3.jpgfinish issues1.jpg

It seems to me that the problem is interpolating both axis simultaneously. I've tried different inserts and different feeds but the profile is coming out very similar which makes me think it is not the control but rather mechanical issue.
In the first picture I noticed a pattern ~0.2mm (~0.01") spacing along the Z-axis which continues onto the radius.

Please advise me on what tests I can do to narrow things down. In writing this I realized that I need to make a finish facing cut and see if there are any patterns there.
The pitch of the screws doesn't seem to coinciding with the ridges that I'm seeing maybe it is a bad bearing somewhere.

Any help and suggestions are highly appreciated.

P.S. now time to read up on how to fit a test scales to see what the axis are doing.
 
It definately looks like vibration to me. Even though the OD finish is acceptable, there are still visual lines there that aren't uniform like normal feed lines would be. I have seen similar finish from a lathe with bad spindle bearings. Throw a test indicator on a part which has been machined so there is no runout, and manually start the spindle at a slow speed (100 rpm or so). Slowly increase the speed and watch the indicator. If the bearings are bad, at some point the indicator will begin to bounce, if not it will continue to run true. I have also seen the same pattern caused by bar whip (which would also make the indicator bounce). If you are using bar stock, be sure that you are using some kind of reduction tube or collar to keep the bar from "whipping" inside the draw tube. The bar stock may also be bent. Have had to cut bars to shorter lengths and/or reduce spindle speed to minimize the vibration.
 
I was literally JUST here and I have lots of bad parts to show for it (long story). Anywho someone much smarter than me told me to switch to a 55 deg diamond with a smaller nose radius (0.008 or so). I was also using a CCGT and he said the IC of the insert was hitting and causing it to deflect enough to ruin the finish. Not hard to try it. I doubted this theory as much as you are now :)
 
Well I did some mechanical testing and was actually secretly hoping that it is just a thrust bearing on the Z-axis as that made sense. The results I believe are acceptable but correct me if I'm wrong. Here is what I tested:
1) Placed a ball bearing into the end of ball screw and read travel in Z direction for Z axis with a test indicator. The maximum I could get by rapidly moving back and forth was about 0.0008". It was the same from both ends and it seemed to have a high spot once per rev of the ball screw.
2) Placed indicator on the base of the machine and moved the turret face up to it to check for total backlash in Z. Set the handwheel to 0.0001" per click, move 20 clicks forward 20 clicks back and it comes back to same zero spot. Actually does the same thing with just one click too.
3) Did the same tests for X axis as #1 and #2 above. The ballscrew end showed no movement and the backlash was also nonexistent.
4) Turned a test piece of stock to have a face and a shoulder for checking spindle runout. Went from 500 rpm to 1200 in both Z and X axis with about the same result of 0.0004"

So nothing jumps out at me except for Z axis thrust bearings seem a bit suspicious. The ridges I'm seeing are much bigger, on the order of magnitude so I'm not sure if any of these runouts are causing them unless there is some dynamic resonance going on.

RobF, I need to get some V and D shape inserts anyway so I'm willing to try your suggestion although I don't understand the problem and what it means "IC hitting"

I might try cutting few different tapers and see if there is a some sort of trend.
 
Well I did some mechanical testing and was actually secretly hoping that it is just a thrust bearing on the Z-axis as that made sense.

Looks like X-axis thrust bearing to me. The OD and ID diameters also look like crap indicating the X axis is bouncing around as Z-axis is feeding in. I had this same problem years ago.

Run each axis in rapid full speed and listen for noise. I will bet X-axis growls a bunch more compared to Z-axis.

Regardless, replace all thrust bearings and then reset backlash parameters if needed. Also, make sure everything is getting lubed correctly.
 
Is it linear rail? If so pull back the way covers and inspect them where the balls ride. When my lathe was crashed it had dimples from the bearings on the rails. I had similar problems.
 
Is the tool sitting too far above (or below) center-line? If the turret is misaligned then the tools are too. If the insert is sitting too high above center-line then it may be doing more pushing than cutting.
 
Hello, I'm working on 1995 Supermax TC-2 cnc lathe. I recently purchased this machine after it has been sitting for around 5 years and the previous owner working on it on and off trying to repair it. It has Mitsubishi Meldas L3 control which I've been getting used to. After repairing few things and getting parameters set correctly I'm doing test parts to see how the lathe performs and ironing out post processor.
The issue I ran into has to do with surface finish on the arc. I'm cutting a part out of 6061 which is short piece of tube ~0.75" long 2.5" o.d. 2"i.d. and has a ball nose end. I'm roughing with CNMG432 for AL and finishing with CCGT 32.51 for AL. The finish pass is 0.01" DOC and 0.0015" IPR. The finish looks good on the straight portion of the tool-path but when the radius is cut the it is not smooth and has steps.
I checked the G-code and it is commanding G03 X.. Y.. R.. so I know it is not cutting in segments. When I first noticed this problem I thought that the CNMG432 was part of the cause but going to CCGT 32.51 still produces similar profile on the radius and improved finish on the straight section.
I think this might be a problem with the machine it self. I decided to take a closer look at the part under the microscope and see if I can learn anything. Below are few pictures from the microscope:
View attachment 85234View attachment 85233View attachment 85232View attachment 85231View attachment 85230

It seems to me that the problem is interpolating both axis simultaneously. I've tried different inserts and different feeds but the profile is coming out very similar which makes me think it is not the control but rather mechanical issue.
In the first picture I noticed a pattern ~0.2mm (~0.01") spacing along the Z-axis which continues onto the radius.

Please advise me on what tests I can do to narrow things down. In writing this I realized that I need to make a finish facing cut and see if there are any patterns there.
The pitch of the screws doesn't seem to coinciding with the ridges that I'm seeing maybe it is a bad bearing somewhere.

Any help and suggestions are highly appreciated.

P.S. now time to read up on how to fit a test scales to see what the axis are doing.


You are almost certainly more experienced than the persons I see making parts that look like this, but when I see this, it's either the program (with out CRC) calls for a sharper nose radius than used, or the operator has the CRC set smaller than that of the actual insert.
 
Run each axis in rapid full speed and listen for noise. I will bet X-axis growls a bunch more compared to Z-axis.

Regardless, replace all thrust bearings and then reset backlash parameters if needed. Also, make sure everything is getting lubed correctly.
The full rapids sound good to me. I will disable the hydraulic pump to cut down on ambient noise and listen again.

Is it linear rail? If so pull back the way covers and inspect them where the balls ride. When my lathe was crashed it had dimples from the bearings on the rails. I had similar problems.
Yes it is. I will go back and look closely at the rails for indentations. However, I should mention that I cut few of these parts at different locations along the Z axis with the same result.

Make sure that lathe is leveled properly and bearing even on all jack screws.
The lathe is leveled properly and is on solid foundation.

Is the tool sitting too far above (or below) center-line? If the turret is misaligned then the tools are too. If the insert is sitting too high above center-line then it may be doing more pushing than cutting.
The turret is aligned since I had to replace coupler gears due to coolant getting into the coupler and rusting it severely. I just rechecked alignment after cutting these parts and it is still good. I need to go back in there and pin the new gears in place.
20130611_025528.jpg20130610_192017.jpg20130722_134518.jpg

I'm using biggest and shortest tools so I know it is not chatter and the tools are cutting very nicely in straight lines.

I think I will pull the Z-axis thrust bearings and inspect them. There are some markings on the tightening nut, so I think someone has "been there" before.
Also, in parameters the X-axis backlash set to 0 and for Z-axis it is set to 20. The manual says it is in counts. Might be worth while to set it to zero and see what 20 counts equals in distance.

Thank you all for your suggestions.
20130624_013057.jpg
 
Before you tear it apart, check the backlash on the actual tool. Stick an indicator to the turret. Run it up to the chuck. Reverse the hand wheel and see how many clicks you get before the indicator moves.

Checking the end of the screw should have been enough, but you never know.
 
Before you tear it apart, check the backlash on the actual tool. Stick an indicator to the turret. Run it up to the chuck. Reverse the hand wheel and see how many clicks you get before the indicator moves.

Checking the end of the screw should have been enough, but you never know.

I did that yesterday and I had no backlash whatsoever. Set the hand-wheel on 0.0001" and even a single click goes out one and comes back to where it started.
 
base your indicator on the chuck then the dial to your turret and get a 2x4 or bar into the turret and try wrenching on it in a few direction and see if your getting any deflections. make sure your in servo lock.
 
I think I might have found a source of the problem.
When turning Z-axis screw by hand I felt little jumps kind of like turning a shaft of servo motor and the rotor jumping between poles. Long story short, the thrust bearings turned out to be fine but the ball nut is contaminated with chips. It is a Ho Fung ball screw and has a two piece nut design with spacer in between to set preload. The screw looks good visually, I didn't see any cracked balls yet. The balls are still shiny but not mirror finish, they have slight hase. There are two sizes of balls in the nut; load and spacer. I will be taking it all apart today and doing cleaning/reassembly. Read some threads on reballing, seems straight forward. Planning to reuse existing balls to see how it goes together and measure backlash. If things look good replace with new balls.
 
So I disassembled the ball nut and cleaned it all out, there were chips in there as suspected. It was actually quite easy to do; just have to keep track of the ball sizes and split them up evenly between the loops. Mine had 66 balls in each loop. The nut/screw feels much better now, nice and smooth. I put the screw back in the machine and performed backlash check. Noticed that the control was over compensating so I reset the backlash to zero and it works perfectly as far as backlash goes.
Re-installing the screw back into the machine I had to put preload the thrust bearing and I noticed that the jumpines (when turning by hand) would increase with more preload so I have replacement thrust bearings coming in today. However this is still not the cause of the ridge problem I believe.
Taking the Z-axis out I had to disconnect the drive timing belt and for the heck of it I spun the servo shaft by hand and noticed that it was jumping between the magnetic poles. I counted 32 steps/poles and at 12mm per turn the steps come out to be around 0.012" which is very close to my ridge size.
Read few threads that suggest that if the servo gets overloaded it can have demagnetized magnets which is how this one feels. So I got a replacement motor on order. Hopefully this is it since it makes perfect sense.
 








 
Back
Top